Retrovirus vectors have become the primary tool for gene delivery in human gene therapy applications (Miller, Nature 357:455-460 (1992)). The ability of retrovirus vectors to deliver an unrearranged, single copy gene into a broad range of rodent, primate and human somatic cells in primary culture makes them well suited for this purpose. Identification and subsequent deletion of the sequences present within retroviral transcripts encoding the packaging signals for avian (E) and murine (.psi.) retroviruses, has enabled development of packaging cell lines to supply in trans the proteins necessary for production of infectious virions, but render the packaging cell lines unable to package their own viral genomic mRNA (Watanabe and Temin, Molec. Cell. Biol. 3(12):2241-2249 (1983); Mann et al., Cell 33:153-159 (1983); and Embretson and Temin, J. Virol. 61(9):2675-2683(1987)). The most important consideration in the construction of retroviral packaging lines has been both the production of high titer vector supernatants free of recombinant replication competent retrovirus, which has been shown to produce T cell lymphomas in rodents (Cloyd et al., J. Exp. Med. 151, 542-552 (1980)) and primates (Donahue et al., J. Exp. Med. 176, 1125-1135 (1992)). Although early murine retroviral packaging lines were highly prone to generation of replication competent retrovirus (RCR) (Cone and Mulligan, Proc. Nat'l. Acad. Sci. USA 81:6349-6353 (1984)) or prone to co-package the .psi.-genome (Mann et al., supra, 1983; Buttimore and Miller, Mol. Cell. Biol. 6(8):2895-2902(1986)), two strategies have evolved for the construction of second generation packaging lines with significantly reduced ability for the generation of RCR. One strategy, embodied by PA317, uses a single genome packaging construct from which the initiation site for second strand synthesis, the 3' LTR, and the .psi. site have been deleted (Miller and Buttimore, Molec. Cell. Biol. 6(8): 2895-2902 (1986)). These modifications eliminate as much as possible homology between the packaging genome and the viral vector to reduce the ability to form recombinants, and have resulted in production of high titer, helper-free virus with many vector systems (Miller and Rosman, BioTechniques 7(9):980-990 (1989)). The second approach has been to divide the packaging functions into two genomes: one that expresses the gag and pol gene products, and the other that expresses the env gene product (Bosselman et al., Molec. Cell. Biol. 7(5):1797-1806 (1987); Markowitz et al., J. Virol. 62(4): 1120-1124 (1988); Danos and Mulligan, Proc. Nat'l. Acad. Sci. (USA) 85:6460-6464 (1988)). This approach eliminated the ability for co-packaging and subsequent transfer of the .psi.-genome, as well as significantly decreased the frequency of recombination due to the presence of three retroviral genomes in the packaging cell that must undergo recombination to produce RCR. In the event recombinants arise, mutations (Danos and Mulligan, supra) or deletions (Boselman et al., supra; and Markowitz et al., supra) within the undesired gene products render recombinants non-functional. In addition, deletion of the 3' LTR on both packaging function constructs further reduces the ability to form functional recombinants. Although early attempts at the generation of two genome packaging lines yielded low titer producer clones (Bosselman et al., supra) producer lines are now available that yield high titer producer clones (Danos and Mulligan, supra; and Markowitz et al., supra).
Packaging lines currently available yield producer clones of sufficient titer to transduce human cells for gene therapy applications and have led to the initiation of human clinical trials (Miller, supra). However, there are two areas in which these lines are deficient. First, design of the appropriate retroviral vectors for particular applications requires the construction and testing of several vector configurations. For example, Belmont et al., Molec. and Cell. Biol. 8(12):5116-5125 (1988), constructed stable producer lines from 16 retroviral vectors in order to identify the vector capable of producing both the highest titer producer and giving optimal expression. Some of the configurations examined included: (1) LTR driven expression vs. an internal promoter; (2) selection of an internal promoter derived from a viral or a cellular gene; and (3) whether a selectable marker was incorporated in the construct. A packaging system that would enable rapid, high-titer virus production without the need to generate stable producer lines would be highly advantageous in that it would save approximately two months required for the identification of high titer producer clones derived from several constructs.
Second, compared to NIH 3T3 cells, the infection efficiency of primary cultures of mammalian somatic cells with a high titer amphotropic retrovirus producer varies considerably. The transduction efficiency of mouse myoblasts (Dhawan et al., Science 254:1509-1512(1991) or rat capillary endothelial cells (Yao et. al., Proc. Natl. Acad. Sci. USA 88:8101-8105 (1991)) was shown to be approximately equal to that of NIH 3T3 cells, whereas the transduction efficiency of canine hepatocytes (Armentano et. al., Proc. Natl. Acad. Sci. USA 87:6141-6145 (1990)) was only 25% of that found in NIH 3T3 cells. Primary human tumor-infiltrating lymphocytes ("TILs"), human CD4+ and CD8+ T cells isolated from peripheral blood lymphocytes, and primate long-term reconstituting hematopoietic stem cells, represent an extreme example of low transduction efficiency compared to NIH 3T3 cells. Purified human CD4+ and CD8+ T Cells have been reported on one occasion to be infected to levels of 6%-9% with supernatants from stable producer clones (Morecki et al., Cancer Immunol. Immunother. 32:342-352 (1991)), and primate or human long-term reconstituting hematopoietic stem cells have only been infected to .ltoreq.1% with a producer of titer of 10.sup.6 per ml on NIH 3T3 cells (van Beusechem et al., Proc. Natl. Acad. Sci. USA 89:7640-7644 (1992); and Donahue et al., supra). If the retrovirus vector contains the neo.sup.R gene, populations that are highly enriched for transduced cells can be obtained by selection in G418. However, selectable marker expression has been shown to have deleterious effects on long-term gene expression in vivo in hematopoietic stem cells (Apperly et.al. Blood 78:310-317(1991)).
An approach that yields significantly increased transduction of mammalian cells in primary culture would be highly advantageous, and this need is currently unmet.